1. Field of the Invention
The present invention relates to an imaging apparatus.
2. Description of the Related Art
With respect to a pixel reading circuit of an image sensor, there has been proposed an image sensor which is equipped with a so-called column-parallel AD (analog-to-digital) converter having an AD converter for each pixel column. In the column-parallel AD converter of this type, a pixel signal and a ramped reference signal are compared with each other by a comparator, and a time from a start of conversion to inversion of an output from the comparator is counted by a counter, whereby AD conversion is performed. Furthermore, in the column-parallel AD converter of this type, the low-speed AD converter is provided for each pixel column, and thus the AD conversion is performed in parallel by the provided low-speed AD converters. For this reason, since the bands of the comparators and an amplifier which constitute the column-parallel AD converter can be lowered, an AD conversion system which uses the column-parallel AD converter has an advantage over other AD conversion systems in terms of an S/N (signal-to-noise) ratio.
Incidentally, a voltage of the reference signal is changed in synchronization with an operation of the counter which counts the conversion time. Consequently, for example, when an eight-bit AD converter is used, a processing time corresponding to 28 steps, i.e., 256 steps, is necessary. Further, when the gradient of the ramped reference signal is increased, the voltage value of the AD converter for each step is increased. That is, there is a method of extending the dynamic range of the AD converter by lowering (roughening) the resolution thereof.
Moreover, as a method of extending the dynamic range without deteriorating a resolution characteristic, there has been known the method which is described in Japanese Patent Application Laid-Open No. 2008-098722. That is, in Japanese Patent Application Laid-Open No. 2008-098722, the reference voltages of which the gradients of the slopes are respectively different are used, the output voltage of the unit pixel and the reference voltage are compared with each other, and further the reference voltage and the second reference voltage are compared with each other. Thus, since the counting operation is performed by using the two comparators, the high-speed AD conversion can be achieved, whereby the dynamic range can be extended. However, in this method, since the two comparators which mainly occupy the AD conversion constitution are necessary for each column, there is a problem that the circuit area resultingly increases.
In the column-parallel AD converter equipped image sensor (that is, the image sensor which is equipped with the column-parallel AD converter), the number of counts of the conversion time is given as the result of the AD conversion. For this reason, it is difficult to extend the dynamic range due to restrictions of the conversion time.
In the general column-parallel AD converter, it is possible to extend the dynamic range by increasing the gradient of the slope of the reference voltage. However, in such a case, the voltage for one LSB becomes large, whereby there is a problem that a trade-off occurs between the voltage and the resolution. For this reason, when the low-illuminance pixels and the high-illuminance pixels mixedly exist in the same row, there is a problem that it is difficult to select an appropriate gradient of the slope of the reference signal. Here, as a method of extending the dynamic range of the AD converter while maintaining the resolution, there are several methods, which include a method of adjusting the gain of the column amplifier for each conversion, a method of performing an addition process in the frame direction, and a method of using the plurality of comparators. However, in any of these methods, there are several problems. That is, the overall circuit constitution becomes complex, the number of frames decreases, and the chip itself becomes large in size. Due to such reasons as described above, in an imaging apparatus which uses the column-parallel AD converter, it is necessary to properly change the dynamic range by using the reference signals having the different gradients in the row without complicating the circuit constitution.